analiza aiqbench
This research advances biomedical NLP by demonstrating that zero-data LoRA adapter merging can effectively transfer medical translation capabilities to low-resource Arabic-script languages using high-resource pivots like Persian. While the method yields significant gains for closely related tongues such as Dari and Urdu, it highlights critical limitations when structural distance between source and target languages is too great for clinical deployment. The study provides a practical framework for scaling healthcare AI in linguistically diverse regions without requiring extensive new datasets for every specific language variant.
Biomedical Machine Translation for Low-Resource Arabic-Script Languages via Cross-Lingual Transfer and LoRA Adapter Merging
arXiv:2607.22300v1 Announce Type: new Abstract: We present a systematic study of healthcare-domain cross-lingual transfer to address the scarcity of biomedical NMT resources for Arabic-script languages. We use Arabic and Persian as higher-resource pivots to improve translation for \textbf{four severely low-resource} targets: Dari (Afghan Persian, a standardised variety of Persian), Pashto, Sorani Kurdish (Central Kurdish, a major standardized variety of Kurdish), and Urdu (closely related to Hindi). Using LoRA fine-tuning on small decoder-only LLMs, we train \textit{domain-specific pivot adapters} and evaluate \textbf{three transfer strategies}: few-shot in-context learning, minimal supervised adaptation, and, to the best of our knowledge, for the first time in this setting, zero-data LoRA adapter merging. Supervised adaptation with just 500 sentences achieves near pivot-language quality for Dari (CHrF++ 41.01) and meaningful gains for Urdu (28.88), while adapter merging reaches within 3.5 CHrF++ of supervised adaptation for Dari at zero additional cost. Pashto and Sorani Kurdish remain insufficient for high-stakes clinical deployment exposing the limits of cross-lingual transfer when structural distance from the pivots is too great. LoRA adapter merging works surprisingly well for closely related languages, even without target-language biomedical data.
Abstract:We present a systematic study of healthcare-domain cross-lingual transfer to address the scarcity of biomedical NMT resources for Arabic-script languages. We use Arabic and Persian as higher-resource pivots to improve translation for \textbf{four severely low-resource} targets: Dari (Afghan Persian, a standardised variety of Persian), Pashto, Sorani Kurdish (Central Kurdish, a major standardized variety of Kurdish), and Urdu (closely related to Hindi). Using LoRA fine-tuning on small decoder-only LLMs, we train \textit{domain-specific pivot adapters} and evaluate \textbf{three transfer strategies}: few-shot in-context learning, minimal supervised adaptation, and, to the best of our knowledge, for the first time in this setting, zero-data LoRA adapter merging. Supervised adaptation with just 500 sentences achieves near pivot-language quality for Dari (CHrF++ 41.01) and meaningful gains for Urdu (28.88), while adapter merging reaches within 3.5 CHrF++ of supervised adaptation for Dari at zero additional cost. Pashto and Sorani Kurdish remain insufficient for high-stakes clinical deployment exposing the limits of cross-lingual transfer when structural distance from the pivots is too great. LoRA adapter merging works surprisingly well for closely related languages, even without target-language biomedical data.
Computation and Language (cs.CL)
arXiv:2607.22300 [cs.CL]
(or arXiv:2607.22300v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2607.22300 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 273 cuvinte
Subiecte conexe
Evaluating Large Language Models on Urdu Idioms
arXiv · 4 sept. 2026
E-CONAN (Entailment, CONtradition And Neutral) Benchmarks: Arabic Textual Entailment and Natural Inference Datasets
arXiv · 12 sept. 2026
Building Large-Scale English-Romanian Literary Translation Resources with Open Models
arXiv · 29 iul. 2026
Links #3: 2026/06 Part 1
LessWrong · 15 iun. 2026